Step 1. Create your algorithm
Create an algorithm you want to use in your workflows.
Quickstart
Step 2. Create a custom block manifest
Create a block manifest that defines and updates custom block metadata.
Step 3. Set up data transfer and storage
Set up the data.json file and additional images to be saved to the /tmp/output directory.
Step 4. Test your block locally
Test custom blocks locally before publishing.
Step 5. Publish your custom block
Publish the block so that it can be used in analytics workflows.
Troubleshooting
How to develop a custom block in Python?
While developing in Python, use the utilities package that contains base classes for processing blocks.
Will other users have access to my algorithm?
No. Other users won't be able to see or download your custom block, as well as access any of the internals.